Spatiotemporal behaviour of a male wild boar, sus scrofa, in the pre-Alps of the canton Fribourg.
Projet for the course “Patterns and trends in environmental data”
Vanina Heinrich, heinrvan@students.zhaw.ch
Abstract
In September 2019, a male wild boar was captured in the pre-Alps of Fribourg. It was equipped with a GPS collar and its movements could be followed for almost 2 whole years. The tracking of this wild boar was analyzed according to 4 aspects: global analysis of its movements, analysis of important points of presence, analysis of the use of the land cover and finally analysis of its activity patterns. The results show a great variability both in the territory occupied and in the use of the different covers. Hunting shows an important influence on its general activity. The results give a direction to future analyses that will be carried out on other individuals in the pre-Alps in order to confirm the elements found during this project.
No studies have been conducted on the movements of wild boar in the Pre-Alps. These analyses are done in the framework of a pre-study. The data are the property of the state of Fribourg and must be treated confidentially.
Introduction and research aims
Wild boars, sus scrofa, are common mammals in Europe (Suter & Fischer, 2021). The study of their movements and habits is used in particular within the framework of the management of their populations but also for the prevention of damage caused to agricultural crops and epidemiology management (Pelayo et al. 2014). Studies carried out within the framework of the EUROBOAR network initiative have shown that wild boars show a great behavioral plasticity and a great adaptability to their environment. This characteristic makes the study of this species complicated and implies that it must be studied in all environments where it is possible to meet it (Brivio et al. 2017). Many studies have been done to understand the spatio-temporal behavior of wild boars but none has been published on wild boars in the Pre-Alps. The pre-Alps present a particular territory due to their steep nature and their landscape coverage.
In this study, the movements of a male wild boar located in the Fribourg pre-Alps (Switzerland) were analyzed over almost two years. The first goal of these analyses was to define the animal’s home range and its use of it. The second goal was to analyze its habits and the effect of an anthropogenic disturbance, hunting, on its habits. As Ohashi and collegues did in their study in 2013, the intensity of indirect daily disturbance caused by human activities was considered as a constant element. As hunting occurs only some months of the year and its directed toward wild animals, it’s considered here as an temporal element that can influence the boar’s habits.
The study of this data set was done from a global analysis to a more detailed analysis. Four elements guided the analysis with the following questions:
Analysis of the general home range: where is the boar located? What are the movements it has made and when? Are there differences between seasons in its occupation of space?
Analysis of areas of high density of presence: did the boar stay in certain places for a long time or did it go back and forth? What is present in these areas?
Analysis of surface cover: what is found in this boar’s home range? How did it use the surface cover? What is found in areas of high density presence?
Analysis of its habits: when is it most active? Are there differences in the use of ground cover? What is the impact of hunting on his habits?
Due to the hilly topography and high elevations, wild boar are expected to move to higher elevations on average in the summer months and then down in the cooler months with potential snowfall. Areas with high density of location points are assumed to be particular sources of food for this boar such as corn fields for example. Concerning the use of the landcover by this boar, the hypothesis is an important use of forest spaces, especially during the day. Finally, in view of the region where the boar is located, which is an area with many trails and anthropogenic activities, it is assumed that the boar presents mainly a nocturnal behavior. Hunting should have an influence on its movements and its use of the landcover.
Material and Methods
Spatiotemporal analysis of the data in this study was performed with R-Studio version 2022.02.3 and ArcGis Pro version 2.8.3.
Data set
The data set used in this project comes from the Forest and Nature Department of the canton of Fribourg. The 3rd september 2019, a male wild boar was captured in the Fribourg pre-Alps and fitted with a GPS collar. This collar transmitted the position of the animal every hour from 6pm to 6am and then once at noon. The collar was programmed by the Nature and Forest departement based on the fact that the wild boar can be predominantly nocturnal in areas with anthropic disturbance (Ohashi et al., 2013 ; Podgorski et al., 2013) Location data could thus be obtained until July 2021. The animal then suddenly disappeared and so did the GPS transmission.
The data set contains 9431 lines, each corresponding to a location defined by X and Y coordinates according to the Swiss coordinate system (LV95/LV03, crs = 2056). The base file contains the following columns:
- Date: date of the survey
- Time: time of the survey
- Altitude: altitude of the GPS point
- X: X coordinate
- Y: Y coordinate
Preprocessing
This data set required some additional columns derived from the base file data to allow for further analysis to answer the questions outlined in the introduction to this project. A first visual analysis of the file in excel was done to observe the data.
The analysis of the “Altitude” column revealed values that could not correspond to the region where the boar was located. Indeed, the highest point in the region is at 2014m (Dent de Lys) and some location points exceeded this value. These points were removed. The same was done for the location points below 500m altitude.
Fig 1: Map of the region
The dates were analyzed and a subclassification by months and years was done. These data were added as new columns to the available data. Date and time were also merged together in a new column. a “Timelag” column was created which indicates the time elapsed between two locations, based on the “Datetime” column created during the previous manipulation of the “Date” and “Time” data. A “Daytime” column has been added. This one is based on the “Timelag” column. If the time difference is less than 65 minutes, it concerns measurements taken at night (from 6pm to 6am), if this difference is greater than 360 minutes, it concerns measurements taken during the day (between 6am and 6pm). The column therefore contains the information Day or Night.
In order to calculate the activity pattern of this boar, I decided to base my calculations on its speed. The speed is calculated by the distance traveled in a certain time lapse. A “steplength” column was created. This one indicates the distance between two rows, i.e. two location points. Finally, a “Speed” column has been created. This one is the result of the “Steplength” column divided by the “Timelag” column.
Finally, to analyze the influence of hunting on the activity of this boar, two new columns were created. The first one is the hunting pressure which was evaluated according to the number of species hunted at the same time on a given date :
Wild boar
2019 - 01.09 until 31.12
2020 - 01.09 until 31.12
Roe deer
2019 - 16.09 until 12.10
2020 - 16.09 until 12.10
Chamois
2019 - 16.09 until 28.09
2020 - 21.09 until 03.10
Deer
2019 - 14.10 until 26.10 and 02.11 until 16.11
2020 - 19.10 until 31.10 and 07.11 until 21.11
This information was provided by the Forest and Nature Department of the Canton of Fribourg. Based on this information, four categories were established for hunting pressure:
0 - no hunting
1 - only wild boar hunting
2 - hunting of wild boar + 1 other species
3 - wild boar hunting + 2 other species
An additional column was created to indicate whether or not hunting was taking place, based on the hunting pressure (0 = no hunting, 1/2/3 = hunting).
The hunting pressure category 3 corresponds only to two week in the year.
Some analyses could only be done with the year 2020, the only full year of data. This is why a second data_frame was created with only the points of 2020, subset of the original data frame.
Methods
I would like to point out that all of the methods described below can be applied to other similar data sets. These same manipulations will be used to study the movements of other wild boars in this same region in a future study.
Global analysis of the home range
Wild boars are territorial (Suter & Fischer, 2021). In order to understand the extent of the male boar’s territory in this study, a visual analysis of its presence was performed. I performed a mcp derived from the location points of this boar to identify its home range. For this purpose, the function st_convex_hull has been applied. It takes a point cloud as input and draws a convex hull around the most distant vertices. The area was then calculated using the chull.area function and compared to the results obtained on ArcGis Pro for validation. The home range was analysed for all the data from 2019 to 2021
The altitude at which this boar was located is an important element considering the region in which it is located. Therefore, a boxplot analysis was performed to see the altitude the boar was located at as a function of the months in 2020.
Density and path
To visualize the areas where wild boar were most often (high density of location points), an analysis via ArcGis Pro was performed by Kernel Density approach. Following this result, and in order to understand if these areas correspond to areas where the wild boar stayed for a long time or if it came back all the time, an analysis of the path by ggplot with the geom_path function was performed for each month in 2020.
Land cover and use of it
With the help of ArcGis Pro, a geopackage regrouping the data of the departements of agriculture and forests of the canton of Fribourg for the year 2020 has been created. Not all parcels could be assigned to a surface category due to lack of the GIS layers. This was then imported into R-studio. As the GIS layer contained only categories in french or german, a column was added with the translation in English of the different types of surfaces for a better understanding. By analyzing the different types of surfaces and thinking about the behavior that a wild boar could have in them, I decided to create three surface overclasses: grassland areas which includes meadows, pastures and other rather open surfaces, forests and finally the surfaces corresponding to woodlands outside forests (Hedges-, fields- and riverbank woods). To facilitate the analysis of the data, I decided to work only with the main surface categories as well as the surface classes I created. For elements requiring detail such as a particular field, I worked with visualisation in ArcGis Pro. The 2020 wild boar data frame and the geopackage data were then combined into a single data frame for further analysis.
A ggplot was made to visualize the presence of wild boar on the different terrains.
To find out the exact use of surface types by wild boar, the percentage of location points present in each surface type was calculated. This manipulation was then refined to identify the type of surface used according to the time of day and also according to the months.
Activity patterns
The speed of the boar was used as a basis for analyzing its activity rate. The GPS collar was programmed to transmit its position once an hour during the night from 6:00 p.m. to 6:00 a.m. and then once at noon. In order to verify this, two subsets were created: one for the day and one for the night. The average of the two subsets was calculated and then they were compared. For a better visualization, a boxplot was made. A day/night comparison was made by boxplot according to the months in 2020.
Finally, as speed can also depend on the type of surface on which the boar moves (behaviour adapted to land cover), a boxplot analysis of the different speeds according to the type of surface was performed.
Effect of Hunting
Thurfjell et al. demonstrated in 2013 that hunting had an influence on boar movement speed. To understand the influence of hunting on this male boar, a boxplot analysis of travel speeds by hunting pressure was performed.
To analyze whether hunting also induced a change in the boar’s surface type use patterns, the percentage of location points in a surface type as a function of whether hunting occurred was performed.
Results
Home range and vertical migration
This male boar was moving in the region of the Fribourg pre-Alps. Its home range was limited by the freeway to the west and by the Sarine river to the east. Its home range from 2019 to 2021 was 158.97 km2.
Fig 2: Home range of the wild boar from september 2019 to july 2021
The analysis of the altitude at which the boar were located from 2019 to 2021 shows that a vertical migration to higher altitudes took place in spring 2020 until June 2020. From July 2020 onwards, it can be observed that the boar again descended to lower altitudes. For the rest of the year 2020, the boar was located at approximately the same altitudes, In 2021, the boar was located within the same altitude range until July 2021 were i went up to higher altitudes (Figure 3, Table 1).
Fig 3: Altitude were the wild boar was, depending on the months and years
## Month Altitude
## 1 January 1022.837
## 2 February 1053.638
## 3 March 1094.439
## 4 April 1072.824
## 5 May 1118.013
## 6 June 1217.348
## 7 July 1263.582
## 8 August 1017.744
## 9 September 1001.240
## 10 October 1079.662
## 11 November 1067.181
## 12 December 1036.450Table 1 : Mean altitude for each month between 2019 and 2021.
Horizontal migration and land cover use in 2020
The point density analysis shows several areas of high density, including one in the northwest near the highway (Figure 4). The point analysis (Figure 5) of wild boar presence by month does not show that wild boar remained for a long time (minimum one month) in the same area. Path analysis by month confirms this trend, indicating that the boar was moving back and forth within its home range (Figure 6).
Fig 4: Density of presence of the wild boar in 2020
Fig 5: Location points by month in 2020
Fig 6: Path analysis by month in 2020
All the analyses carried out concerning the use of the different types of surface show that the wild boar was mainly located in the forest. Its presence in the forest did not vary much between months. Its presence in summer pastures was more important in May, June and July 2020 than in the other months. In January 2020, about 20% of the wild boar were located in areas not identified in the available geopackage compared to 10% in December 2020 (Figure 7).
Fig 7: Land cover use per month in 2020
The wild boar were primarily located in the forest during the day (between 6:00 am and 6:00 pm) and diversified the areas where they were located at night (Figure 8).
Fig 8: Land cover use depending on daytime for each month in 2020
An analysis on ArcGis Pro of areas with many location points showed that the GIS layer provided by the Departement for agriculture did not identify all areas (Figure 9).
Fig 9: Some location points could not be attributed to a specific land cover with the provided GIS layer
Activity patterns
The speed of the boar does not show a big visual difference from month to month (Figure 10). It averages 0.64 m/min during the day (6am to 6pm) and 5.04 m/min at night (6pm to 6am).
Fig 10: Speed depending on daytime for each month in 2020
The boar moved slower on average in the forest than in other surface types except for litter surfaces (Table 2).
## Landcover Speed
## 1 Forests 4.929045
## 2 Hedges, fields and riverbank woods 6.360040
## 3 Litter surfaces 2.623450
## 4 Permanent meadows 7.151576
## 5 Permanent pastures 7.232080
## 6 Summer pastures 6.431069
## 7 Tree, overlaying areas 8.110733Table 2 : Mean speed for each landcover category in 2020.
Effect of hunting
Boxplot results show greater forest use during the hunting season in 2020 (Figure 11). Grassland use is similar between periods.
Fig 11: Landcover use depending on hunting in 2020
The boxplots show an increase in velocity when hunting pressure is present compared to a period without hunting (Figure 12). The mean of speed increased during the day and during the night (Table 3, Table 4).
Fig 12: Speed depending on the hunting pressure in 2019 and 2020
## Huntpressure Speed
## 1 0 0.6114687
## 2 1 0.5750352
## 3 2 0.8364226
## 4 3 0.6313324Table 3 : Mean speed during day depending on hunting pressure from 2019 to 2021.
## Huntpressure Speed
## 1 0 4.700096
## 2 1 5.067419
## 3 2 6.327295
## 4 3 4.690600Table 4 : Mean speed during night depending on hunting pressure from 2019 to 2021.
Discussion
Home range and vertical migration
The study of the movements of the wild boar in this study allowed a better understanding of its habits, its use of the land cover and the influence of the hunt on its spatial behavior. The data collected with the GPS collar allowed the study of its vertical and horizontal migration as well as the visualization of its home range. The results show that the boar in this study had a territory of 158.67 m2 which is about five times larger than the territories recorded for male boars which range from 4.5 to 30 km2 (Suter & Fischer, 2021). This difference could be explained by the hilly topography of the home range of this boar. Studies on other wild boars in the same region would allow us to establish whether this is an isolated case linked to this individual or a generalization for the Pre-Alps region.
Horizontal migration and land cover use
correlation with the description in Suter and Graf 2021
general trend of wild boar to leave the natural habitats and their margins between march and july (Gerard et al. 1991)
Activity patterns
favour natural habitats more marked during resting times than during activity phases (Gerard et al. 1991)
Clear nocturnal activity due to anthropogenic disturbance (Ohashi et al. 2013 ; Podgorski et al. 2013).
Effect of hunting
nocturnal activity should increase when they are exposed to hunting pressure (Keuling et al. 2008 ; Ohashi et al. 2013). Nocturnal activity did not increase during hunting season (Brivio et al. 2017).
These results were validated by personal visualization and comparison with elements found in publications. Some answers were provided by discussion with the wildlife ranger of the region. A detailed statistical analysis is necessary to be able to affirm if the trends observed on the obtained graphs are significant or not. This analysis will be carried out in a second time, outside the framework of this work.
References
Brivio F., Grignolio S., Brogi R., Benazzi M., Bertolucci C. & Apollonio M. (2017) : An analysis of intrinsic and extrinsic factors affecting the activity of a nocturnal species : The wild boar. Mammalian Biology 84: 73-81.
Gerard J-F., Cargnelutti, Spitz F., Valet G. & Sardin T. (1991) : Habitat use of wild boar in a French agroecosystem from late winter to early summer. Acta theriologica 36 (1-2): 119-129.
Keuling O., Stier N. & Roth M. (2008) : How does hunting influence activity and spatial usage in wild boar Sus scrofa L. ? Eur. J. Wildl. Res 54: 729-737.
Ohashi H. et al. (2013) : Differences in the activity pattern of the wild boar Sus scrofa related to human disturbance. Eur J. Wildl Res 59 : 167-177.
Pelayo A., Quiros-Fernandez F., Casal J. & Vicente J. (2014) : Spatial distribution fo wild boar population abundance : Basic information for spatial epidemilogy and wildlife management. Ecological Indicators 36 : 594-600.
Podgorski T., Bas G., Jedrzejewska B., Sönnichsen L., Sniezko S., Jedrzejewski W. & Okarma H. (2013) : Spatiotemporal behavioral plasticity of wild boar (Sus scrofa) under contrasting conditions of human pressure : primeval forest and metropolitan area. J. Mammal. 94: 109-119.
Suter S.M & Fischer C. (2021) : Wildschwein, p .316-139 in : Graf R.F. & Fischer C. (2021): Atlas der Säugetiere. Schweiz und Liechtenstein. Schweizerische Gesellschaft für Wildtierbiologie SGW, Haupt Verlag, Bern.
Thurfjelle H., Spong G. & Ericsson G. (2013) : Effects of hunting on wild boar Sus scrofa behaviour. Wildlife Biology 19 : 87-93.
R-Code
#Analyzing path in 2020 in a more funny way
p <- ggplot(jojo2020, aes(X, Y)) + geom_path(show.legend = FALSE, mapping = aes(color = "indianred4")) + transition_time(Date) + labs(title = "Date : {frame_time}")
path2020 <- animate(p, duration = 120, render = gifski_renderer())
anim_save("jojo_path2020.gif", path2020)